4.0
@hn_5b14b2
about 1 month ago
Grok2 is right up there in the middle of the pack with the big dogs on a lot of the benchmarks. It's no joke. So I think Grok 3 could be really interesting.
xAI's second-generation LLM, launched in beta on X with Grok-2 and Grok-2 mini; weights later open-sourced in 2025, using dense attention with MoE elements.
@hn_5b14b2
about 1 month ago
Grok2 is right up there in the middle of the pack with the big dogs on a lot of the benchmarks. It's no joke. So I think Grok 3 could be really interesting.
@hn_f0aaa2
about 1 month ago
Grok did OK? (Grok 2 Mini Beta) https://i.imgur.com/mvnusFd.jpeg
@hn_8c5d03
about 1 month ago
Here a write up for issues back with Grok 2, demoing prompt injection from uploaded docs or other user's posts, data leakage, hidden prompt injection, ASCII Smuggling, etc. https://embracethered.com/blog/posts/2024/security-probllms-... The zero-click data leakage was fixed (at least in the dedicated webapp) but xAI never acknowledged that these issues are security vulnerabilities - which was quite an interesting response.
@hn_0d3867
about 1 month ago
It is deeply funny that's the case - it happened with Grok-2 as well - but I can't imagine that remaining the case when they (ostensibly) scale to superintelligence. After all, it would be unwise to build a superintelligence that has both the desire and the means to kill you.
@hn_945c36
about 2 months ago
I'm following AI news and models for few years now and i have not read about your Grok2 controversy. Nonetheless, i do not use grok and i do not try it out due to it being part of Musk. I'm also not aware that Grok2 was communicated as the top model in any relevant timespan at all. Perhaps it just didn't deliver? Or a lot more people are not awaare of how to use it or boycot Musk. After all he clearly doesn't care for any rules or laws it is probably a very high risk sending anything to grok.